Dataframe mean calculation problem: do we have to loop?

[sparkt]

I got it, though I expected something much simpler.

df['C'] = 0.0
for i in df.index:
    df['C'][i] = df[['A', 'B']][i:].mean().mean()
print(df)

But do we have to use this for loop to go through all values? One of the best things of dataframe is that it can deal with a complex frame once only when applying function, without the need of going through looping which is not good for optimal performance.

I hope there's a more elegant solution!